Initial support for ostree based systems #2557
Conversation
|
This is definitely work in progress. I'm looking for suggestions on how we should implement this properly. |
Thanks for looking into this 👍 much appreciated. I made some comments and maybe I understand the ostree better when we meet at oSC :) |
Unfortunately I'm struggling to get funding for travelling to DevConf.CZ so it's unlikely that I'll be able to go to oSC this year :( |
travier
force-pushed
the
main-atomic-ostree-support
branch
from
8d46a4b to
a1f2cb0
Compare
|
This is still hackish but a bit less 😅 |
|
@travier Couple of things here:
|
Will do.
I need help figuring out:
|
|
Friday is usually a good time for me. |
This week and next week I won't be able to make it. I will be traveling to NUE and next week I'm in Portland for a workshop. The week after we can schedule a meeting and Friday is also a good day for me. That would mean October 18th |
|
That works for me too. |
|
@travier @Conan-Kudo Hi, unfortunately I still don't get my voice back and sound like a mouse. I think we need to move the meeting to another day |
Temporary method to place all ostree specific code paths behind a single check. May be converted to configuration option later.
- Do not replace but append to the existing kernel command line as it is correct - Do not fix the path to the kernel & initrd as those are also correct
Do not rebuild the initrd in the ostree case and use the one from the ostree commit instead. It is expected that it includes all the needed modules.
schaefi
force-pushed
the
main-atomic-ostree-support
branch
from
a1f2cb0 to
b93421c
Compare
|
@travier in preparation to the conversation I was so free to rebase your code to the upstream main and fixed the conflict. Hope this is ok ? |
Added a small refactor to prevent code duplication and applied tests for code parts that I believe are good to go. For other code parts I added a "TODO: ostree:" marker and spared out the code from being covered by the coverage checker. For those remaining changes we need to have a conversation how to proceed
| @@ -418,6 +420,25 @@ def setup_live_image_config( | |||
| has_graphics = True | |||
| if 'serial' in self.terminal_output or 'serial' in self.terminal_input: | |||
| has_serial = True | |||
|
|
|||
| # Find the ostree=... parameter from the BLS config and add it to the boot options | |||
| # TODO: ostree: This code looks like it is required for any ostree based boot process and should live in its own scope e.g. an OSTree class | |||
There was a problem hiding this comment.
Open for the Friday conversation
| ] | ||
| ) | ||
| else: # pragma: nocover | ||
| # TODO: ostree: this code should be part of an OSTree class find the existing initrd in the ostree case |
There was a problem hiding this comment.
Open for the Friday conversation
| @@ -90,6 +91,29 @@ def get_kernel( | |||
| filename=kernel_file, | |||
| version=version | |||
| ) | |||
|
|
|||
| if Defaults.is_ostree(self.root_dir): # pragma: nocover | |||
| # TODO: ostree: This code looks similar to the one in builder/live and should imho exist only once in a OSTree class | |||
There was a problem hiding this comment.
Open for the Friday conversation
defaults: Add method to check for ostree case
Temporary method to place all ostree specific code paths behind a single
check. May be converted to configuration option later.
bootloader/config/grub2: Handle ostree case for BLS configs
correct
system/kernel: Find kernel in ostree case
kiwi/builder/live: Handle ostree case
Do not rebuild the initrd in the ostree case and use the one from the
ostree commit instead. It is expected that it includes all the needed
modules.
kiwi/bootloader/config/grub2: Handle ostree case
See: #38